Large area focal plane arrays (“FPAs”) are used in a variety of infrared radiation (“IR”) imaging applications to obtain imaging data having high spatial resolution. An IR imager containing one or more FPAs generates imaging data based on the detection of absorbed IR energy that was emitted or transmitted from a target surface area (scene) facing the FPAs. Large area FPAs are sometimes also incorporated into imaging spectrometer systems to provide both high spatial and high spectral resolution data on an imaged scene.
As is well known in the art, a FPA in a conventional IR imager is in the form of a detector array chip and an associated read-out circuit, and includes a two-dimensional array of pixels arranged in a planar layout on the detector array chip. The pixels correspond to respective geometrical areas on the surface of the detector array chip, and operate as IR radiation detectors that collect the minority charges that are generated in the detector array chip when it absorbs IR radiation supplied from a scene imaged by the FPA. The pixels of the detector array chip collect the minority charges for a predetermined interval, known as an integration time interval, after which the pixels are sampled to ascertain the amount of collected minority charges. After sampling, the pixels are cleared or reset, and then again collect minority charges generated based on the absorption of radiation for another integration time interval. The collection of photogenerated minority charges by the pixels of the detector array chip, and then the sampling and resetting of the pixels by the read-out circuit, are performed iteratively as the FPA images a scene or a sequence of scenes, portion-by-portion.
In many imaging applications, it is desired that an FPA imager detects the absorption of radiation at all portions on the FPA designed to have imaging capabilities, in other words, at all of the pixels of the FPA. If all or substantially all of the pixels of a FPA are operating, the FPA can generate imaging data that represents all or substantially all of the radiation absorbed from the scene being imaged. An imaging apparatus having these operating characteristics is known in the art as a very high operability imager.
Very high operability, long wavelength infrared (“LWIR”) imaging spectrometers, which can provide high spatial and high spectral resolution imaging data on a scene based on absorption of LWIR radiation, are desirable for many applications, such as science, missile defense and space surveillance. For example, a LWIR imaging two-dimensional Fourier transform infrared spectrometer (“FTS”), such as a spectrometer having a cutoff wavelength λc of about 17 μm, having very high operability LWIR FPAs and that can also image a scene over a relatively short time interval, is needed for imaging applications ordinarily performed on a satellite platform. The satellite platform imaging applications, for example, require generation of a sequence of images in synchrony with the movement of the FTS's mirror, where each image is obtained based on the absorption of LWIR radiation by the pixels of the detector array chip, so that high spatial and high spectral resolution atmospheric sounding data can be obtained.
As early prior art IR FPA imagers, which are commonly known as small format FPAs, contained only a small collection of radiation detectors, a large scene was partitioned into multiple portions that the FPA sequentially sampled so that high spatial and high spectral resolution IR imaging data could be obtained. Consequently, the time interval for imaging the entire scene was relatively long, which was undesirable because the resulting imaging data could not accurately capture rapidly changing information contained in the scene. Although small format FPAs can be operated to image larger portions or the entirety of a scene to achieve a shorter interval for imaging the scene, this results in the reduction of spatial resolution, which is undesirable for high resolution imaging applications.
FPAs with increased numbers of pixels, thus, were developed to satisfy the requirements for increased imaging resolution. A preferred, and conventional, prior art method of producing an IR FPA with a large number of pixels is by hybridization, where each pixel of a separately manufactured detector array chip containing a plurality of pixels is coupled by an individual metal column to a silicon read-out integrated circuit (“ROIC”) chip. The ROIC includes electrical circuitry that conditions the photogenerated charges collected at minority charge collectors, which are p/n or n/p junction collection diodes formed at the respective pixels, and that transfers data representative of the minority charges collected at the individual pixels to downstream signal processing electronics. The downstream electronics uses the minority charge data to generate corresponding imaging data with high spatial and/or high spectral resolution, such as is required for atmospheric sounding data.
For ease of reference, the construction of a single pixel included in a detector array chip is described below as having a horizontal portion that has length and width dimensions, and a vertical portion extending orthogonally from the horizontal portion and that corresponds to the thickness or depth of the chip. It is to be understood that, in a detector array chip, pixels are typically arranged in a pattern extending horizontally along both the length and width dimensions, where the pattern of the pixels usually is dictated by the specific imaging application. FIG. 1A is a cross-sectional (depth) view of an exemplary prior art pixel 10 having a conventional radiation detector architecture, and FIG. 1B is a top down plan (horizontal) view of the pixel 10 showing only the detector components. FIG. 1A is a cross-sectional view of the pixel 10 taken along line 1A-1A in FIG. 1B. Referring to FIG. 1B, the pixel 10 corresponds to a horizontal portion of a detector array chip (not shown) having a width W and a length L, where W and L are conventional values and can be the same value, such as 120 μm. Referring to FIG. 1A, the pixel 10 includes an implant 12 formed in a base optical energy absorber layer 14 made of p-type or n-type semiconductor material, such as HgCdTe. The absorber layer 14 is formed on a substrate 16, such as CdZnTe. The implant 12 forms a p/n or n/p junction diode by converting a portion of the semiconducting absorber layer 14 to n-type or p-type, respectively. The absorber layer 14 and the substrate 16 extend horizontally on the detector array chip so as to be co-extensive with the horizontal portion of the detector array chip corresponding to the pixel 10. The implant 12 typically extends horizontally over a large fraction, for example 90%, of the horizontal portion of the detector array chip corresponding to the pixel 10. The p/n or n/p junction formed by the implant 12 and absorber layer 14 constitutes a collection diode that functions as the radiation detector for the pixel 10.
Still referring to FIG. 1A, as is well known in the art, in operation IR radiation usually enters the pixel 10 through the substrate 16. The radiation is absorbed in the absorber layer 14, which results in the generation of minority and majority charges. The minority charges diffuse toward and are collected at the p/n or n/p junction that constitutes the collection diode of the pixel 10. As the horizontal portion of the detector array chip over which the implant 12 extends for the pixel 10 is almost co-extensive with the horizontal portion of the detector array chip over which the absorber layer 14 in the pixel 10 extends, there is a high likelihood that the collection diode of the pixel 10, if operable, will collect all or substantially all of the minority charges induced by the IR radiation absorbed by the pixel 10.
Referring again to FIG. 1A, a metal layer 20 overlies the implant 12 and electrically connects the collection diode of the pixel 10 to a metal pad 22. The metal pad 22, which can be an indium bump, facilitates the transfer of the photogenerated minority charges collected at the collection diode to processing components in the ROIC chip portion (not shown) of the FPA imager. In addition, an insulating protective layer 18 overlies portions of the absorber layer 14 that do not contain the implant 12 to avoid shorting of p/n junctions.
Based on the current state of IR FPA technology, when a LWIR detector array chip including the pixels 10 is manufactured using the required p-type and n-type semiconductor materials according to conventional chip manufacturing processes, a significant fraction, typically about 5%-10%, of the collection diodes is inoperable, such that the corresponding pixels 10 are inoperable. Defects created in a detector array chip during deposition of the absorber layer 14 or the manufacture of the detector array chip cause some of the implants 12, and thus the corresponding collection diodes, to be inoperable. If a defect in the detector array chip intersects the p/n or n/p junction forming the collection diode of a pixel, the defect degrades or shorts the p/n or n/p junction. When a collection diode is degraded or shorted, dark current in the collection diode overwhelms the current due to photogenerated minority charges from the absorber layer 14 of the pixel 10, such that the collection diode cannot provide any photogenerated minority charge information for the pixel 10 of the FPA from which imaging data can be derived.
Referring again to FIG. 1B, the implant 12 extends over about 90% of the horizontal portion of the chip corresponding to the pixel 10. As the collection diode for the pixel 10 occupies a very large part of the horizontal portion of the detector array chip corresponding to the pixel 10, there is a high likelihood that the collection diode on the pixel 10, namely the implant 12, will intersect a manufacturing defect in the detector array chip p/n junction area, such that the current collected at the collection diode of the pixel 10 that is based on photogenerated minority charges would be overwhelmed by dark current in the collection diode. Where several of the pixels of an FPA are not operable because their respective collection diodes intersect detector array chip defects, as will likely occur when the FPA includes pixels having the detector architecture of the pixel 10 illustrated in FIGS. 1A and 1B, the FPA cannot generate imaging data that faithfully represents all or substantially all of the radiation supplied from the imaged scene and absorbed in the pixels. In other words, an FPA including inoperable pixels does not produce imaging data representative of the portions of a scene to which the pixels including the inoperable collection diodes correspond. Consequently, an FPA chip including an array of the pixels 10 likely would not have the desired high operability, and thus high spectral and high spatial resolution, that the FPA chip theoretically can achieve, because one or more of the pixels likely would not be operable due to expected manufacturing defects formed in the detector array chip material.
Conventional detector array chip processing techniques cannot be modified to eliminate the creation of defects in the detector array chip during manufacture thereof, without expending substantial additional effort and incurring substantial additional expense. It is known that one cause of the manufacturing defects expected to form in an LWIR detector array chip is a natural result of the growth of absorber materials, such as HgCdTe, on a substrate, such as CdZnTe, because the lattice of the absorber materials does not perfectly match the lattice of the substrate. Although the formation of some defects during the detector array chip growth process can be reduced, it has been found that very costly modifications of the standard detector array chip manufacturing processes achieve only a relatively small percentage reduction in the number of defects formed in LWIR detector array chips. Based on the expected formation of defects in a detector array chip, LWIR detector array chips having many thousands of the pixels 10, and also having very high operability, such as about 99.9% or greater, are nearly impossible to manufacture using conventional chip processing.
Research and development efforts have focused on increasing the operability of LWIR FPAs in view of expected defects in the detector array chip. According to one prior art technique, each of the pixels of a hybrid LWIR FPA including an array of the pixels 10, such as a detector array chip having a 256.times.256 array of the pixels 10, becomes a sub-element of a super-pixel. A super-pixel is formed from four adjoining pixels 10, where each of the pixels includes a single p/n or n/p junction collection diode. See Stobie, James A. et al., SPIE Vol. 4818, pg. 213 (2002), incorporated by reference herein. As the pixels 10 in the FPA chip become sub-elements of a super-pixel, the probability that a sub-element of a super-pixel has a defect is lower than the probability that the single, large super-pixel has a defect. For example, if a pixel is divided into four sub-elements, the probability of any sub-element having a defect is lowered by a factor of four relative to the probability that a single, large pixel has a defect. The ROIC is designed to allow the collection of photogenerated minority charges only from non-degraded or non-shorted sub-elements for each super-pixel. The degraded or shorted sub-elements do not contribute to the overall minority charge collection of the pixel. A disadvantage of this prior art technique, however, is that a loss of signal results for every sub-element that is shorted or degraded. For example, if two sub-elements of a super-pixel have their respective collection diodes intersecting a defect such that the collection diodes are shorter or degraded, the photogenerated minority charges are not collected at the regions of these sub-elements so that as much as 50% of the energy incident on the super-pixel can be lost and, therefore, not accounted for in the imaging data.
In an alternative prior art approach for improving FPA operability, a detector array chip is manufactured using pixels having the lateral collection diode (“LCD”) architecture. See H. Holloway, 49 J. Appl. Phys., pg. 4264 (1978), incorporated by reference herein. FIG. 2A is a cross-sectional view of an exemplary prior art pixel 50 having an LCD radiation detector architecture, and FIG. 2B is a top down plan view of the pixel 50 showing only the p/n junction diode components. FIG. 2A is a cross-sectional view of the pixel 50 taken along line 2A-2A in FIG. 2B. Referring to FIG. 2B, the exemplary pixel 50 is shown extending over the same horizontal portion of a detector array chip as the pixel 10 of FIG. 1B to allow for ease of comparison between the detector architectures of the pixels 10 and 50. Referring to FIGS. 2A and 2B, the pixel 50 includes a plurality of p/n collection diodes and, referring to FIGS. 1B and 2B, the horizontal portion of the detector array chip over which each of the implants 12 extends, and thus the horizontally extending p/n junction area of each of the collection diodes, is much reduced in relation to the horizontal portion corresponding to the implant 12 of the pixel 10. Each of the implants 12 in the pixel 50 has a circularly-shaped horizontal portion. In the typical implementation of the pixel 50, the pixel 50 is included in a detector array chip intended for LWIR application, each of the implants 12 has a radius of about 10 μm or less and the W and L dimensions of the pixel 50 range from about 40 μm to about 120 μm. Referring again to FIG. 2B, each of the implants 12 is a sub-element of the pixel 50 and the implants 12 are horizontally spaced apart from one another a distance equal to no more than the diffusion length for the minority charge carriers in the pixel 50. The pixel 50 otherwise has substantially the same construction as the pixel 10, except that the insulating protective layer 18 is interposed between metal contact layer 20 and portions of the absorber layer 14 that do not contain implants 12 to avoid shorting the p-side to the n-side of the p/n junctions. The metal pad 22 overlies the metal contacts 20 and connects the collection diodes in the pixel 50 to one another to form a single, large radiation detector. In operation, the sub-elements of the pixel 50 collect substantially all of the minority charges photogenerated in the absorber layer 14, including those photogenerated between the implants 12. See D'Souza, A. I. et al., 29 J. Electron. Mater., 630 (2000), incorporated by reference herein.
Referring to FIGS. 1B and 2B, the implants 12 of the respective collection diodes in the pixel 50 extend over a much smaller horizontal portion, typically about 10%, of the detector array chip than the horizontal portion of the detector array chip corresponding to the implant 12 in the pixel 10. As a result, the probability that a defect in the semiconductor material of the pixel 50 portion of the detector array chip would intersect the p/n junctions in the pixel 50, so as to short or degrade the collection diodes therein, is reduced by about a factor of 10 relative to the probability that the collection diode of the pixel 10 would intersect a defect. See Wijewarnasuriya, P. S. et al., 28 J. Electron. Mater., p. 649 (1999), incorporated by reference herein.
Although the detector architecture of the pixel 50 improves operability of a FPA including the pixels 50, the guarantee of very high operability, where all pixels in a detector array chip consisting of the pixels 50 are fully operational, however, cannot be made. Defects in the detector array chip still are likely to intersect collection diodes in the pixels 50 and, thus, result in shorted or degraded collection diodes in the pixels 50 for which photogenerated minority charges are overwhelmed by dark current.
It is also known in the art to use a microlens in conjunction with a pixel of an FPA to reduce the size of the horizontal portion of the pixel over which the p/n junction of the collector diode extends. FIG. 3A is a cross-sectional view of an exemplary prior art pixel 100 that includes a microlens 32, and FIG. 3B is a top down plan view of the pixel 100 showing only the p/n junction diode component. FIG. 3A is a cross-sectional view of the pixel 100 taken along line 3A-3A in FIG. 3B. Referring to FIG. 3B, the pixel 100 is shown extending over the same horizontal portion of a detector array chip as the pixels 10 and 50, as shown in FIGS. 1B and 2B, to allow for ease of comparison among the detector architectures of the pixels 10, 50 and 100. Referring to FIGS. 1A, 1B, 3A and 3B, the pixel 100 is similar in construction to the pixel 10, except that a single, smaller sized implant 12 is at the center of the pixel 100 and the microlens 32 covers the horizontal portion of the absorber 14 through which optical energy will enter the pixel 100. Referring to FIG. 3B, the implant 12 in the pixel 100 has a circularly-shaped horizontal portion having a radius of about 30 μm and the pixel 100 preferably has W and L values near or about equal to 120 μm. Like the conventional and LCD detector architectures of the pixels 10 and 50 respectively described above, the use of the microlens 32 in the pixel 100 does not guarantee very high operability for a large detector array consisting of the pixels 100.
Therefore, a need exists for a pixel for use in an IR detector array chip, where the pixel has a detector architecture which provides substantially one-hundred percent operability of the pixels in a large array, where conventional detector array chip processing can be used to manufacture a detector array chip including such high operability pixels and where an FPA imager including a detector array chip formed with such high operability pixels can be sampled at conventional FPA imager sampling rates.